virus expression (xenotropic viruses/f/st and BIOF mice)

Transcription

1 Proc. Nati Acad. Sci. USA Vol. 8, pp , January 1983 Genetics H-2-linked regulation of xenotropic murine leukemia virus expression (xenotropic viruses/f/st and BIOF mice) ROBERT A. YETTER, JANET W. HARTLEY, AND HERBERT C. MORSE III* Laboratory of Microbial Immunity and Laboratory of Viral Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 225 Communicated by Wallace P. Rowe, October 25, 1982 ABSTRACT A high proportion of lymphocytes from F/St mice produce infectious xenotropic murine leukemia virus (X- MuLV) and express high levels of cell surface antigens, termed XenCSA, related to the major glycoprotein of X-MuLV. In crosses of F/St with AKR, the high-virus phenotype of F/St was found to be recessive and was shown to be governed by a single locus, Cxv-1, less than 2 centimorgans from H-2K. The close association of Cxv-1 with the H-2 complex was confirmed by the observation that BIO.F mice, congeneic for the H-2 region of F/St, expressed high levels of infectious X-MuLV and XenCSA, whereas C57BL/ 1 mice and other C57BL/1 H-2 congeneic strains did not. Studies of hybrid mice homozygous for Cxv-1s, but segregating for a chromosome 1 X-MuLV induction locus (V locus) of F/St, demonstrated that the high-virus phenotype of F/St was dependent on the interaction between Cxv-1 and the chromosome 1 V locus. Genetic information coding for xenotropic murine leukemia viruses (X-MuLVs) is present in multiple copies in the cellular DNA of all mice (1, 2), but different inbred strains vary greatly in their expression of this information as infectious virus or as virus-coded glycoprotein gp7o (XenCSA, X-MuLV-related cell surface antigens) on the surface of lymphocytes (3, 4). Earlier studies demonstrated that some ofthe differences in expression ofinfectious X-MuLV could be attributed to the presence ofloci for induction ofinfectious virus (V loci) in some strains and their absence from others (5-8). For many strains, the X-MuLV loci have been shown to be present on chromosome 1 at or near Bxv- 1 (5, 6, 9). Variations in XenCSA levels among inbred strains were also shown to be influenced by V loci (8), but other differences were attributable to the effects of a locus on chromosome 4 linked to Fv-1 (1, 11). Recently, we described another locus that has a major effect on expression of X-MuLV as infectious virus and XenCSA in F/St mice (9). A high proportion ofthymocytes and spleen cells from mice of this strain produced infectious ecotropic MuLV and X-MuLV and expressed high levels of XenCSA. Studies of hybrids between F/St and other inbred strains showed that the high ecotropic MuLV phenotype off/st was governed by three or more dominant, independently assorting V loci. However, the high X-MuLV phenotype of F/St was found to be a recessive trait that was governed by a single locus. This locus was not linked to the X-MuLV V locus of F/St on chromosome 1 or to marker loci on chromosomes 4, 5, 6, 7, 9, or 14. Expression of infectious ecotropic MuLV was not detectably affected by this locus. In this report, we demonstrate that the locus governing X- MuLV expression in F/St mice is located on chromosome 17 within or tightly linked to the major histocompatibility complex The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisenent" in accordance with 18 U. S. C solely to indicate this fact. 55 (H-2). This locus appears to regulate the levels to which infectious X-MuLV and XenCSA are expressed from loci on chromosome 1. MATERIALS AND METHODS Mice. F/St (H-2P, Fv-l('r), Ly-9.2) mice were bred in our colony. B1. F/Sg (H-2P, Fvlb) mice were from a colony established from breeders provided by Jack Stimpfling. C57BL/ 1SnJ (H-2b, FV-lb), AKR/J (H_2k, Fv-jn), BIO.BR/SgSnJ (H- 2k, Fvlb), and BDP/J (H-2P, Fv-l, Ly-9.1) mice were purchased from The Jackson Laboratory. P/N (H-2P, Fv-ln) mice were from the colonies of the National Institutes of Health. Antibodies. Monoclonal anti-h-2kk was purchased from Becton Dickinson. Fluorescein-labeled monoclonal anti-ly-9. 1 was prepared from supernatants of cultures of the 3C7 cell line, generously provided by Jeff Ledbetter (12). Fluorescein-labeled Fab fragments of goat anti-mouse IgG2 were generously provided by B. J. Fowlkes, National Institute of Allergy and Infectious Diseases. The reagent for assayingxencsa consisted of fluorescein-labeled F(ab')2 fragments of IgG obtained from the serum ofa rabbit (R283) immunized with SIRC cells infected with NZB X-MuLV, prepared as described (4). These antibodies react preferentially with cell surface gp7s related to the envelope gp7s ofx-mulv and mink cell focus-inducing MuLV. Virus Assays. Suspensions of lymphocytes from thymus and spleen were prepared for infectious center assays of mitomycin C-treated (25 Ag/ml for 3 min at 37C) cells as described (4, 13). X-MuLV was assayed for focus induction in sarcoma viruspositive, leukemia virus-negative (S+L-) mink lung cells (14); titers are expressed as focus-forming units (ffu) per 17 lymphoid cells. Ecotropic MuLV was assayed by plaque induction in XC cells as described (13); titers are expressed as plaque-forming units (pfu) per 17 cells. Expression of XenCSA was assayed by flow microfluorometry (FMF) using a fluorescence-activated cell sorter (FACS II; Becton Dickinson FACS Systems) and fluorescein-labeled antibodies from rabbit R283 (4). The intensity ofstaining detected on a population of lymphocytes is expressed as the mean fluorescence value (4) and was determined by using either linear or logarithmic amplification of signals from the photomultiplier tubes. XenCSA values are given as geometric means. Assays for H-2Kk and Ly Single-cell suspensions from spleen (1,ul, 2 x 17 cells per ml) were allowed to react with unlabeled or labeled monoclonal antibodies for 3 min at 4 C. Abbreviations: MuLV, murine leukemia virus; X-MuLV, xenotropic MuLV; V locus, a locus governing inducibility of MuLV; S+L-, sarcoma virus-positive, leukemia virus-negative; XenCSA, X-MuLV-related cell surface antigens; ffu, focus-forming units; pfu, plaque-forming units; FMF, flow microfluorometry. * To whom reprint requests should be addressed at: LMI-NIAID-INIH, Building 7, Room 335, Bethesda, MD 225.

2 al'vf%6 Genetics: Yetter et al Proc. Nati. Acad. Sci. USA 8 (1983) After two washes, cells that had been allowed to react with unlabeled reagents were incubated for an additional 3 min at 4C with fluorescein-labeled anti-mouse IgG2. Individual animals were designated as being homozygous, heterozygous, or negative for alleles at these loci by comparing the staining detected on their spleen cells by FMF with that detected on positive, negative, and F1 controls. Table 1. Q ax C.) wcd I" RESULTS Expression of X-MuLV by mice from crosses between AKR and F/St. Previous studies of F1 mice from crosses between F/St and strains with low spontaneous expression of X-MuLV, such as AKR, showed that the high X-MuLV phenotype of F/St was recessive, whereas the high ecotropic MuLV phenotype was dominant (9). Further analyses of F1 x F/St mice demonstrated that the high X-MuLV phenotype of this strain was determined by a single recessive gene for which no chromosomal location was established (9). In the present study, crosses between AKR and F/St mice were tested for possible linkage between alleles at the locus governing X-MuLV expression and alleles at H-2. Spleen cells and thymocytes from (AKR X F)F1 X F/St mice were assayed for infectious X-MuLV by cocultivation with mink S+L- cells. Cells from the same preparations were also assayed for XenCSA and alleles at H-2K by FMF. The results of these tests demonstrated the following points: First, as in earlier studies of (NFS X F)F1 X F/St and (C57L X F)F1 X F/St mice (9), the high X-MuLV phenotype of F/ St was controlled by a single recessive gene with greater effects on thymocytes than on spleen cells (Table 1, Fig. 1). A high proportion of thymocytes from 23 of 53 (AKR X F)F1 X F/St mice (43%) produced infectious X-MuLV and expressed high levels of XenCSA (mean fluorescence values greater than 4; Fig. 1A). In contrast, thymocytes from the remaining 3 mice produced little or no infectious X-MuLV and expressed low levels of XenCSA. Similar studies of spleen cells from 29 of these mice showed several differences from the studies of thymocytes. First, the proportions of virus producing cells formed a near continuum from < 1-3/17 cells to 13.9/17 cells instead ofgiving two distinct sets. In addition, XenCSA levels on spleen cells from mice with a high proportion of virus-producing cells were not significantly higher than XenCSA levels on spleen cells from mice producing little or no infectious X-MuLV. Second, studies of the same mice for segregation of alleles at H-2K demonstrated that the high-virus phenotype for thy- - CD 5, 4, 3, 1 2, 1,I 1, c X-MuLV, log1o(ffu/17 cells) FIG. 1. Thymocytes from 53 (AKR x F)F1 x F/St mice (A) and spleen cells from 29 of the same mice (B) were tested for the proportions of virus-producing cells (x axis) and expression of XenCSA (y axis) and for alleles at H-2K. Values for H-2KP/H-2KP mice are given as o; values for H-2KI/H-2KP mice are given as *. XenCSA values for thymocytes and spleen cells of all heterozygotes at H-2k were less than 46. mocytes (infectious X-MuLV and XenCSA) was observed only in mice homozygous for the H-2K allele of F/St [H-2P (15)], whereas thymocytes from H-2Kk/H-2KW heterozygotes, except for one mouse, produced no infectious virus and expressed low levels of XenCSA (Fig. 1A). In studies of spleen cells, the majority of mice with low proportions of virus-producing spleen cells (<11/17 cells) were heterozygous at H-2K, whereas spleen cells from mice homozygous at this locus generally had higher proportions of virus-producing cells (Fig. 1B). Finally, thymocytes from backcross segregants ofboth the high and low X-MuLV phenotypes produced equivalent levels of infectious ecotropic MuLV (data not shown). Expression of infectious X-MuLV and XenCSA by lymphocytes Ecotropic MuLV, X-MuLV (mink S'L-), XenCSA, log1o(pfu/17 cells)* loglo(ffu/17 cells)* mean fluorescencet Mice Thymus Spleen Thymus Spleen Thymus Spleen F/St , AKR (AKR x F)F, C57BL/ B1O.BR BlO.F , (BlO.BR x B1O.F)F P BDP (BDP x F)F , (P x F)F ,64 28 * Values indicate the mean for two to nine mice 2-5 months of age. The SEM was less than 1. (log1o) for all strains and Fjs tested. t Fluorescence values determined by using linear amplification of signals from the photomultiplier tubes. Values indicate the mean for 2 to 15 mice 2-6 months of age. The SEM was less than 22% of the mean for all strains and Fjs tested. A _~~~~~~~~~~~ * o. soo....~~~~~ 8

3 Genetics: Yetter et al Proc. Natl. Acad. Sci. USA 8 (1983) 57 ~2: z1 ~ ~~ 1 1 1, Fluorescence intensity FIG. 2. Fluorescence profiles of thymocytes from F/St (-), B1O.BR (---), and BlO.F mice (-.-) that had reacted with fluoresceinlabeled rabbit antibodies to XenCSA. The profiles were generated by using logarithmic amplification of signals from the photomultiplier tubes. These results demonstrate that, in crosses between AKR and F/St, a single locus, which we will call Cxv-1, governs high versus low expression of infectious X-MuLV and XenCSA by thymocytes and, to a lesser extent, high versus low expression of infectious X-MuLV by spleen cells. The locus has no discernible effect on ecotropic MuLV expression. F/St carries a recessive, permissive allele (Cxv-l ) at this locus, whereas AKR carries a dominant, restrictive allele (Cxv-1l. The data also indicate that Cxv-1 is located less than 1.9 ± 1.8 centimorgans from H-2K. Expression of X-MuLV in B1O.F Mice. If Cxv-1 were losely linked to or within H-2, it would be expected that mice bearing the same H-2 haplotype as F/St might express high levels ofinfectious X-MuLV and XenCSA. To evaluate this possibility, thymocytes and spleen cells from mice of the H-2 congeneic strains B1O.F and B1O.BR as well as C57BL/1 were assayed for these two parameters ofvirus expression; C57BL/1 and the H-2 congeneics carry a X-MuLV V locus at Bxv-1 (5). The results of these assays (Table 1) demonstrated that, in comparison to lymphocytes from C57BL/1 and B1O.BR mice, a high proportion of thymocytes and spleen cells from BlO.F mice produced infectious X-MuLV, equivalent to the levels observed with F/St. In addition, XenCSA levels on thymocytes (Fig. 2, Table 1) and spleen cells (Table 1) were in the range of values observed with cells from F/St. It should be noted that infectious ecotropic MuLV was also produced by a high proportion of thy- Table 2. mocytes and spleen cells from B1. F mice. All isolates tested have been B-tropic in host range and the high ecotropic MuLV phenotype of this strain appears to be due to maternal transfer of virus, at least in part through the milk (unpublished observations). Earlier studies have shown that certain other C57BL/ 1 H-2 congeneic strains express high levels of ecotropic MuLV as a result of non-h-2-linked differences from C57BL/1 (16). To verify that the differences in X-MuLV expression between B1O.F and B1O.BR mice were determined by a H-2-linked locus, thymocytes from (B1O.BR x BlO.F)F1, (B1O.BR X B1.F)F1 X F/St, and F/St X (B1.BR X B1.F)F1 mice were examined for production of infectious X-MuLV, expression of XenCSA, and segregation of alleles at H-2K. As for F/St, the high-virus phenotype was recessive in the F1 (Table 1); the proportion of virus-producing cells was governed by a single locus (Table 2); and the locus determining the high- or low-virus phenotype was closely linked to H-2K (Table 2). Expression of X-MuLV in Crosses Between F/St and P or BDP Mice. The H-2P haplotype, carried by F/St and BlO.F, is also borne by strains BDP and P, which are closely related to one another but are not related to F/St. In contrast to F/ St and BlO. F, these two strains spontaneously produce no infectious X-MuLV and express only low levels of XenCSA on lymphocytes (Table 1). Further, spleen and tail fibroblast cells could not be induced to produce infectious X-MuLV after treatment with Salmonella lipopolysaccharide, 5-iododeoxyuridine, or both (data not shown); this is also in contrast to F/St (9). F1 hybrids between F/St and BDP or P expressed the high X- MuLV and XenCSA levels characteristic of the F/St parent (Table 1), indicating that high virus expression requires H-2P homozygosity (or homozygosity at a closely linked gene) and an inducible xenotropic V locus. This was further tested by examining the progeny ofthe cross BDP X (BDP X F)F1, in which mice with a chromosome 1 V locus from F/St should be high for X-MuLV expression whereas backcross segregants homozygous for chromosome 1 from BDP should be negative. As shown in Table 3, of 27 BDP X (BDP X F)F1 mice, 14 animals expressed high levels of infectious X-MuLV and XenCSA in thymus, indicating that the high-virus phenotype was controlled by a single gene. These mice were also tested for segregation of alleles at Ly-9, a chromosome 1 linkage marker located 2 centimorgans from the X-MuLV V locus of F/St (9). In 22 of these 27 mice, the virus phenotype assorted in association with alleles at Ly-9 such that the majority of BDP-type mice (homozygous for Ly-9.1) were low for virus expression, whereas the majority of the heterozygotes at this locus were Expression of infectious X-MuLV and XenCSA by thymocytes of mice from crosses between F/St and (B1O.BR x B1O.F)Fj F/St x (B1O.BR x B1O.F)Fj (B1O.BR x B1O.F)Fj x F/St k/p* p/p* k/p* pp* X-MuLVt XenCSAt X-MuLVt XenCSAt X-MuLVt XenCSAM X-MuLVt XenCSAt , * Alleles at H-2K determined by FMF using monoclonal anti-h-2kk. t Values indicate loglo(ffu/17 cells) on mink S'L- cells. Virus titers >2. (log1o) are considered to be significantly increased. t Mean fluorescence values determined by using logarithmic amplification of signals from the photomultiplier tubes.

4 `8 Genetics: Yetter et al Table 3. Expression of infectious X-MuLV and XenCSA by thymocytes of BDP x (BDP x F)F1 mice Ly 9.1/Ly 9.2* Ly 9.1/Ly 9.1* X-MuLVt XenCSAt X-MuLVt XenCSAt 4.2 1, , , , , , < < < , , < , ,35 * Alleles at Ly-9 were determined by using monoclonal anti-ly-9.1. t Values indicate log1o(ffu/17 cells) on mink S'L- cells. Virus titers >2. (loglo) are considered to be significantly increased. * Mean fluorescence values determined by using logarithmic amplification of signals from the photomultiplier tubes. Apparent recombinants between alleles at Ly-9 and the F/St X- MuLV V locus. high. This gives a linkage estimate of 18.5 ± 7.5, which is compatible with the distance between Ly-9 and the X-MuLV V locus of F/St as determined by iododeoxyuridine induction studies. DISCUSSION The data presented here demonstrate that high expression of X-MuLV by lymphocytes of F/St mice is dependent on the interactions of two loci: Cxv-), closely linked to H-2 on chromosome 17, and a X-MuLV V locus on chromosome 1. Alternative alleles at Cxv-1 appear to regulate the extent to which chromosome 1 X-MuLV V loci are expressed as infectious virus and as XenCSA. The close association between Cxv-1 and H-2K was documented by the observation that among 76 backcross and outcross mice involving F/St or B1. F and mice carrying a chromosome 1 X-MuLV V locus [i.e., AKR or B1.BR, which carry Bxv-1 (5)] only mice homozygous for H-2K! had a high proportion of thymocytes that produced infectious X-MuLV and expressed high levels of XenCSA (Fig. 1, Table 2). These results also showed that the allele at Cxv-l associated with H-2KY was recessive and permissive for X-MuLV expression, whereas that linked to H-2Kk in two H-2k haplotypes of independent origin was dominant and restrictive for X-MuLV expression. The finding that lymphocytes from the H-2P strains BDP and P produced no infectious X-MuLV and expressed low levels of XenCSA suggested that homozygosity at H-2P was not sufficient to produce the high-x-mulv phenotype expressed by F/St. The necessity for a dominant MuLV V locus was indicated by the observation that a high proportion ofthymocytes from (BDP X F)F1 and (P X F)F1 mice (homozygous for H-2P but heterozygous for the chromosome 1 X-MuLV V locus of F/St) produced infectious X-MuLV and expressed high levels of XenCSA (Table 1). Further, linkage analyses of BDP X (BDP X F)F1 backeross mice showed assortment of the high X-MuLV phenotype of F/St in association with the chromosome 1 locus Ly- 9 (Table 3). Proc. Natl. Acad. Sci. USA 8 (1983) The view that the alternative alleles at Cxv-) determine the extent to which lymphocytes produce infectious X-MuLV or express XenCSA from chromosome 1 X-MuLV V loci is further supported by the studies of C57BL/1 congeneic resistant mice. The C57BL/1 strain has a single X-MuLV V locus, Bxv- 1 (5), on chromosome 1. However, thymocytes from C57BL/ 1, BlO. BR (Table 1), B1O.A, or BlO. S mice (unpublished observations) produce little or no infectious X-MuLV and express low levels of XenCSA, whereas thymocytes from H-2P BlO.F (Table 1, Fig. 2) produce high levels of infectious virus and express high levels of XenCSA. The effect of Cxv-) on expression of infectious X-MuLV and XenCSA is greater on thymocytes than on spleen cells; the proportions of virus-producing cells in thymuses of Cxv-V homozygotes are generally 1- to 1-fold higher than in the spleens ofthe same animals (Fig. 1; Table 1; ref. 9). Preliminary studies comparing infectious X-MuLV production by thymocytes and separated splenic T and B cells indicate that little of this difference is due to the smaller proportion of T cells in spleen than in thymus (unpublished dat. The effects of Cxv-) on XenCSA expression by thymocytes and spleen cells appear to be related to its effect on infectious X-MuLV. Assays of thymocytes from three sets of backcross mice-(akr X F)Fj X F/St (Fig. 1), (NSF X F)F1 X F/St, and (C57L X F)F1 X F/St (ref. 11)- showed that XenCSA values were detectably increased only when greater than.1% of the cells were producing infectious virus. Because the proportions ofvirus-producing cells in spleen rarely exceed this level it is perhaps not surprising that XenCSA values for splenocytes are not increased. The means by which Cxv-l exerts its effect on X-MuLV expression is not known, but both intracellular and extracellular mechanisms can be envisioned. Because ofthe close association with the H-2 region, we have carried out preliminary tests for indications that Cxv-) acts as an immune response gene determining responsiveness to X-MuLV env gene products. Samples of SIRC cells infected with NZB X-MuLV were allowed to react with sera from 15 (AKR x F)F1 X F/St mice (eight high-virus and seven low-virus mice) and then exposed to fluorescein-labeled anti-mouse F(ab')2. FMF assays of the cells gave uniformly low fluorescence levels with sera from both high- and low-virus mice, in tests in which standard positive mouse and rabbit sera gave high levels of fluorescence. Although these results suggest that the low-virus and low-xencsa phenotype is not simply a function of X-MuLV immune responsiveness, the possibility of absorption of crossreactive antibodies in mice that all express high levels of ecotropic virus has not been ruled out. Further testing of immunological control mechanisms, both humoral and cellular, is indicated, as well as investigations of control at the level of transcription or translation of the viral genome. It will also be of interest to determine the relationship of Cxv-) to other retrovirus and leukemia regulatory genes known to be within or closely associated with the H-2 complex region (17). It should be noted that the high-expression MuLV locus of NZB mice, Nzv-l, is distinct from the xenotropic virus induction and regulation loci described here, because it is not located on chromosome 1 (C. A. Kozak, personal communication) nor is it linked to the H-2 complex (18). There was no indication of Cxv-l-mediated regulation of endogenous ecotropic MuLV expression in the H-2P crosses studied here and in a previous report (9). However, in contrast to X-MuLV, for which endogenous expression without spread of virus is measured, these ecotropic virus phenotypes reflect the amplification by exogenous infection characteristic ofthe highly inducible multiple V loci of AKR and F/St or of maternally transmitted virus as in the case of B1. F. An H-2P mouse strain

5 Genetics: Yetter et al carrying a single ecotropic V locus on the restrictive Fv-lb background, such as might result from Caesarian derivation of BlO. F, could provide an opportunity to determine whether Cxv-1 affects expression of both ecotropic MuLV and X-MuLV or X-MuLV only. Finally, it should be noted that of the B1. F mice scored as expressing high levels of X-MuLV and XenCSA a high proportion were also positive for virus of the dual-tropic recombinant MCF class (19). The significance of this finding in relationship to the mechanism of Cxv-1 regulating activity and the mechanism of generation of MCF viruses is not yet known. The authors gratefully-acknowledge the technical contribution of Ms. E. Rudikoff, Ms. N. Wolford, Ms. J. Chandler, and Mr. C. Shifler and the assistance of Ms. S. Grove, Ms. V. Shaw, and the National Institute of Allergy and Infectious Diseases editorial staff in the preparation of this manuscript. We also thank Drs. W. P. Rowe, W. F. Davidson, and C. A. Kozak for their numerous helpful discussions. R.A.Y. is a recipient of a fellowship from the Cancer Research Institute of New York. 1. Buckler, C. E., Hoggan, M. D., Chan, H. W., Sears, J. F., Kahn, A. S., Moore, J. L., Hartley, J. W., Rowe, W. P. & Martin, M. A. (1982)J. Virol 41, Hoggan, M. D., Buckler, C. E., Sears, J. F., Chan, H. W., Rowe, W. P. & Martin, M. A. (1982)J. Virot 43, Morse, H. C., III, & Hartley, J. W. (1982) in Current Topics in Microbiology and Immunology, eds. Henle, W., Hofschneider, P. H., Koprowski, H., Melchers, F., Rott, R., Schweiger, H. G. & Vogt, P. K. (Springer, Berlin), Vol. 98, pp Proc. Natl. Acad. Sci. USA 8 (1983) Morse, H. C., III, Chused, T. M., Boehm-Truitt, M., Mathieson, B. J., Sharrow, S. O. & Hartley, J. W. (1979) J. Immunol. 122, Kozak, C. & Rowe, W. P. (1978) Science 199, Kozak, C. A. & Rowe, W.* P. (198)J. Exp. Med. 152, Datta, S. K. & Schwartz, R. S. (1976) Nature (London) 263, Chused, T. M. & Morse, H. C., III (1978) in Origins of Inbred Mice, ed. Morse, H. C., III (Academic, New York), p Morse, H. C., III, Kozak, C. A., Yetter, R. A. & Hartley, J. W. (1982) J. Virol. 43, Morse, H. C., III, Chused, T. M., Hartley, J. W., Mathieson, B. J., Sharrow, S.. & Taylor, B. A. (1979)J. Exp. Med. 149, Morse, H. C., III, Taylor, B. A., Kozak, C. A., Chused, T. M., Sharrow, S. O., Hartley, J. W. & Stockert, E. (1982)J. Immunol. 128, Ledbetter, J. A., Goding, J. W., Tsu, T. T. & Herzenberg, L. A. (1979) Immunogenetics 8, Kawashima, K., Ikeda, H., Hartley, J. W., Stockert, E., Rowe, W. P. & Old, L. J. (1976) Proc. Natl Acad. Sci. USA 73, Peebles, P. R. (1975) Virology 67, Beisel, K. W., Cullen, S. E., Savarirayan, S., Stimpfling, J. H. & David, C. S. (1978) Immunogenetics 7, Datta, S. K., Tsichlis, P., Schwartz, R. S., Chattophadhyay, S. K. & Melief, C. J. M. (1978) Immunogenetics 7, Steeves, R. & Lilly, F. (1977) Annu. Rev. Genet. 11, Datta, S. K. & Schwartz, R. S. (1977) Virology 83, Hartley, J. W., Wolford, N. K., Old, L. J. & Rowe, W. P. (1977) Proc. Natl Acad. Sci. USA 74,